There exists the case that a data processing device is brought into an abnormal operation due to bugs or failures in a circuit, bugs in a program, data irregularly input from an external device, or the like. In order to determine the cause of the abnormal operation, it is preferably to record and analyze histories such as an internal state of the data processing device, a signal between internal circuits, or an input/output signal from/to an external device. Also in adjustment for performance improvements or the like, it is effective to record and analyze the histories in order to grasp the availability ratio or the like of the internal circuit or the input/output signal.
Usually, historical data is recorded on a memory for recording the history with a recording time. The historical data is then read out from the memory and analyzed, when the abnormal operation occurs.
Here, as a technique relevant to historical-data collection, there has been a conventional technique that collects each of two kinds of log information; that is, detailed log information for investigating the cause of a failure when the failure occurs, and normal log information for grasping a normal state. Furthermore, there has been a conventional technique that respectively records identifiers corresponding to various parameters in the operation process log indicating the event performed in an application group or the internal state so as to generate an identifier recording/operation process log, and records the identifier recording/operation process log on a log recording memory. In addition, there has been a conventional technique that records the event occurrence frequency of the same event pattern developed in the event information when an analysis object program is executed.
Patent Document 1: Japanese Laid-open Patent Publication No. 2006-172401
Patent Document 2: Japanese Laid-open Patent Publication No. 2005-275942
Patent Document 3: Japanese Laid-open Patent Publication No. 2009-223487
The memory for recording the histories has a finite capacity and hence, there is no other choice but the data processing device has to stop history recording or overwrite sequentially from the head of the memory when the memory becomes full. Therefore, there has been a problem in that information helpful for specifying the cause of the occurrence of the abnormal operation is sometimes unrecorded or deleted by overwriting.